Sanction teardown · D. Kansas, USA · 2026-02-02
Lexos Media IP, LLC v. Overstock.com, Inc.
What happened
In D. Kansas, USA, a filing relied on ChatGPT to help draft legal argument. The court identified the following problems with the citations in that filing:
-
False Quotes (Case Law)Court determined Lexos attributed language to Baldwin Graphic Systems that does not appear in that opinion (quotation fabricated).
-
False Quotes (Case Law)Court found an explanatory parenthetical quoted from AstraZeneca did not exist in that opinion (quotation fabricated).
-
Misrepresented (Case Law)Court found citation to AVM Techs. LLC v. Intel Corp. was incorrect (wrong reporter/reference) and that the cited authority does not support the proposition asserted.
-
False Quotes (Case Law)Court found Lexos misattributed a quotation to Rodriguez-Felix that actually appears in Daubert and noted Rodriguez-Felix does not contain the quoted language.
-
False Quotes (Case Law)Court found Lexos quoted Microsoft (i4i v. Microsoft) in support of a point about incomplete claim constructions, but the quoted language does not exist in that opinion.
-
False Quotes (Case Law)Court found Lexos attributed specific language and holdings to Cordis that do not appear in Cordis and noted the patents discussed did not contain the quoted language (misquotation/nonexistent quotation).
-
Misrepresented (Case Law)Court found several page-specific citations (e.g., Flexuspine, cited at 879 F.3d 1369, 1375) did not support the proposition asserted and the cited page lacked the discussed admissibility analysis (misrepresented authority).
-
Fabricated (Case Law)Court found a cited decision 'Hockett v. City of Topeka, No. 19-4037-DDC, 2020 WL 6796766, at *3' does not exist; quotation and parenthetical attributed to it are fabricated.
-
False Quotes (Case Law)Court found the quoted parenthetical attributed to Liquid Dynamics is not in the opinion (quotation fabricated).
Which AI tool
ChatGPT. Note: Charlotin's public database records tool attribution only where a court order, brief, or reporting on the matter states it explicitly; "unidentified" or "implied" means the record indicates AI use but does not name a specific product — we do not guess.
Outcome
Brief struck, monetary sanction, firm compliance (monetary penalty: 12000 USD.)
Additional detail
Order to show cause was here. Defendant later obtained summary judgment, in part because plaintiff failed to cite any legal authority (once the hallucinated material was stricken).
How Citation Safe would have caught this
Citation Safe runs three deterministic layers before a brief is filed: (1) does the citation exist against CourtListener's database of published opinions, (2) if quoted, does that exact language appear in the source, (3) does the cited case actually support the proposition it is cited for. Fabricated case citations fail Layer 1. Fabricated or misattributed quotations fail Layer 2 even when the underlying case is real. Misrepresented holdings — a real case cited for a proposition it does not support — are the target of Layer 3. None of these checks involve asking another language model whether the citation looks right; they are lookups and text-matches against the actual source, which is why a hallucinated citation has to survive a direct lookup against the authoritative source — not another model's opinion — to earn a VERIFIED stamp; our measured false-verify rate is published live at /quality.
Check a brief before you file it → · See our live false-verify rate
Source: https://www.damiencharlotin.com/documents/1438/Lexos_v._Overstock_USA_2_February_2026.pdf, via Damien Charlotin's public AI Hallucination Cases Database (CC0).